Ultraviolet absorption cross-sections of hot carbon dioxide
-
Upload
independent -
Category
Documents
-
view
1 -
download
0
Transcript of Ultraviolet absorption cross-sections of hot carbon dioxide
www.elsevier.com/locate/cplett
Chemical Physics Letters 399 (2004) 490–495
Ultraviolet absorption cross-sections of hot carbon dioxide
Matthew A. Oehlschlaeger *, David F. Davidson, Jay B. Jeffries, Ronald K. Hanson
High Temperature Gasdynamics Laboratory, Department of Mechanical Engineering, Stanford University, Stanford, CA 94305-3032, USA
Received 7 September 2004; in final form 11 October 2004
Available online 2 November 2004
Abstract
The temperature-dependent ultraviolet absorption cross-section for CO2 has been measured in shock-heated gases between 1500
and 4500 K at 216.5, 244, 266, and 306 nm. Continuous-wave lasers provide the spectral brightness to enable precise time-resolved
measurements with the microsecond time-response needed to monitor thermal decomposition of CO2 at temperatures above 3000 K.
The photophysics of the highly temperature dependent cross-section is discussed. The new data allows the extension of CO2 absorp-
tion-based temperature sensing methods to higher temperatures, such as those found in behind detonation waves.
� 2004 Elsevier B.V. All rights reserved.
1. Introduction
Accurate ultraviolet (UV) absorption cross-sections
for CO2 at high-temperatures are needed for many
applications including: combustion and flow diagnos-
tics, photolytic conversion of CO2 to CO, and the
modeling of planetary atmospheres. Many previous
studies [1] have focused on the absorption in the vac-
uum UV by room temperature CO2 but few have
investigated the absorption at elevated temperature orfor wavelengths longer than 200 nm. At longer UV
wavelengths absorption by hot CO2 occurs primarily
due to pre-dissociative transitions in the CO2ð1B2Þ CO2ð1Rþg Þ electronic system [2,3]. However, small con-
tributions to the cross-section may also come from
transitions to the 1A2,3A2, and
3B2 states [4,5]. Here
we present new absorption cross-section measurements
made behind reflected shock waves at high tempera-tures (1500–4500 K) utilizing time-resolved absorption
of four selected continuous-wave (cw) lasers at 216.5,
244, 266, and 306 nm. Recently, Schulz et al. [6] have
shown that the UV CO2 absorption feature is spectrally
0009-2614/$ - see front matter � 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2004.10.060
* Corresponding author. Fax: +1 650 723 1748.
E-mail address: [email protected] (M.A. Oehlschlaeger).
smooth at wavelengths longer than 190 nm. The
continuous nature of the absorption and its strongwavelength- and temperature-dependence allow for
temperature sensing strategies using two laser wave-
lengths as demonstrated by Jeffries et al. [7]. This study
extends the absorption cross-section measurements of
Schulz et al. [6] to the higher temperatures needed to
study the gas temperature behind detonation waves.
Previous studies of the CO2 absorption cross-section
at high temperatures in the range from 190 to 355 nmhave utilized shock tubes [2,6,8,9], furnaces [3,10], and
laminar flames [11]. The earliest of these studies, carried
out by Generalov et al. [8], reported the cross-section at
238 and 300 nm measured behind shock waves at tem-
peratures of 1400–6300 K, but their cross-section results
are uncertain because of difficulty in correcting for time-
dependent CO2 dissociation with the instrumentation of
the era. Ibreighith and Roth [9] reported cross-sectionmeasurements at 214.7 and 236 nm made behind shock
waves at temperatures of 1400–2800 K and Koshi et al.
[2] have reported cross-sections at 193 nm measured
behind shock waves at temperatures of 1500–2700 K.
Hartinger et al. [10] have measured cross-sections at
193 nm in an electrically heated furnace at 573–1273
K and Jensen et al. [3] have published UV cross-sections
(230–350 nm) at 1523–2273 K also measured in a
M.A. Oehlschlaeger et al. / Chemical Physics Letters 399 (2004) 490–495 491
furnace. Joutsenoja et al. [11] measured the UV absorp-
tion cross-section of CO2 in a laminar pre-mixed flame
across the wavelength range of 200–270 nm and temper-
ature range 1400–1700 K. Schulz et al. [6] reported the
absorption spectrum of shock-heated CO2 between 200
and 320 nm from 900 to 3050 K. Most recently laser-induced fluorescence from CO2 was observed following
UV absorption [12,13].
The strong temperature-dependence of the absorp-
tion cross-section has enabled new UV absorption-based
temperature measurement strategies in combustion
gases [7,14]. A CO2 thermometry diagnostic [7,14]
requires precise absorption data over an extended tem-
perature range and this need motivates this work. Addi-tionally, the temperature-dependence of the absorption
cross-section measurements enables us to speculate on
the photophysics of CO2 UV absorption.
This work follows up on previous CO2 absorption
cross-section measurements made in our laboratory
[6] in which light from a deuterium lamp was trans-
mitted through shock-heated gas mixtures and dis-
persed with a small imaging monochromator; thetime-resolved transmission spectrum (10 ls time reso-
lution) was recorded with a modified frame transfer
CCD camera (kinetic spectrograph). These measure-
ments provided the UV CO2 absorption spectrum
from 900 to 3050 K, found that this absorption fea-
ture is spectrally smooth, and suggested use of the
strong wavelength- and temperature-dependence of
the feature for high-temperature sensing strategies.However, the limited spectral brightness of the deute-
rium lamp limited the signal-to-noise ratio and
required millisecond temporal averaging. Because of
this time averaging, the previous work above 2300 K
required corrections for thermal decomposition of
CO2, and these corrections limited the maximum tem-
perature to 3050 K. The current experiments were per-
formed using light from cw lasers at selectedwavelengths detected with UV photodiodes with a
wide dynamic range. The increased spectral brightness
and detection sensitivity enables microsecond time res-
olution with a tenfold increase in signal-to-noise. This
improved time resolution enables direct monitoring of
the thermal decomposition of the CO2, and thus
allows an extension of the temperature range from
3000 to 4500 K. The absorption cross-section at thesevery high temperatures is of interest for temperature
sensing behind detonation waves [14] where many
other diagnostic techniques utilizing discrete absorp-
tion transitions become difficult because of collisional
broadening at the high pressures. In addition, the
time-resolution of the experiments enables a determi-
nation of the absorption cross-section immediately fol-
lowing the passage of the shock wave, hence, none ofthe data presented here require chemical model-based
corrections for thermal decomposition.
2. Experimental
Mixtures of 2%, 5%, and 10% CO2 dilute in argon
were studied behind reflected shock waves, which pro-
vide stationary gas mixtures at known initial tempera-
ture and pressure. Experiments were performed in ahigh-purity pressure-driven stainless steel shock tube
(14.13 cm diameter) that has been described elsewhere
[15]. The initial pressure and temperature of the post-
shock experimental conditions were calculated from
the ideal shock wave relations and the measured incident
shock speed. The pressure and temperature were taken
to be the vibrationally equilibrated conditions. As dis-
cussed below, the vibrational relaxation time is veryshort (�1–5 ls depending on temperature and pressure)
behind the reflected shock wave in these experiments.
The vibrationally relaxed experimental temperature
and pressure have uncertainties of 0.6% and 1.1%,
respectively [16]. The gas samples were mixed external
to the shock tube in a stainless steel tank with an inter-
nal stirring system with research grade gases (99.998%
CO2 and 99.999% Ar). Absorption measurements weremade at a location 2 cm from the endwall of the shock
tube through UV-grade fused silica windows. The re-
flected shock conditions ranged from 1500 to 4500 K
and 40 to 170 kPa.
CW laser radiation was generated at four different
wavelengths: 216.5 nm (2–3 mW) was generated by dou-
bling the output of an Ar+ pumped dye-laser, operating
at 433 nm, in a BBO external frequency-doubling cavity,previously described in our methyl radical work [15]; 306
nm (2–3 mW) was generated by intra-cavity frequency
doubling in a Nd:YVO4-pumped (532 nm) dye laser
operating at 612 nm, as described in our OH radical
studies [17]; and 244 and 266 nm (1.5 mW each) were
produced by the single pass of a focused laser beam at
488 nm (Ar+ line) or 532 nm (Nd:YVO4) through
angle-tuned BBO. The harmonics (244 and 266 nm)were separated from the fundamental beams (488 and
532 nm) with a set of Pellin-Broca prisms.
The UV laser beams were split into two components:
one, less than 1 mm in diameter, passing through the
shock tube to be absorbed by CO2(I), and one detected
prior to absorption as a reference (I0). These two beams
were detected using amplified S1722–02 Hamamatsu sil-
icon photodiodes (risetime <1.0 ls, 4.1 mm diameter)and recorded on a digital oscilloscope. The absorption
cross-sections were determined using Beer�s law
I=I0 ¼ expð�rðk; T ÞnLÞ;where I is the transmitted laser intensity, I0 is the refer-
ence beam intensity, r(k,T) [cm2 molecule�1] is the
wavelength- and temperature-dependent CO2 absorp-tion cross-section, n [molecules cm�3] is the CO2 number
density, and L is the absorption path length (diameter of
the shock tube 14.13 cm). For these experiments the
Fig. 2. Example CO2 absorbance (ln(I0/I)) at 216.5 nm for experiment
with thermal decomposition. Initial mixture: 2% CO2/Ar. Vibrationally
equilibrated incident shock conditions: 1812 K and 16.3 kPa. Initial
(prior to decomposition) vibrationally equilibrated reflected shock
conditions: 3838 K, 83.1 kPa. Note the vibrational relaxation after
incident shock-heating and the incubation period prior to decompo-
sition after reflected shock-heating.
492 M.A. Oehlschlaeger et al. / Chemical Physics Letters 399 (2004) 490–495
minimum detectable absorbance (ln(I0/I)) was <0.1%
and absorption cross-sections were not derived for
experiments in which the absorbance was less than 0.7%.
The absorbance (ln(I0/I)) traces for several experi-
ments are shown in Fig. 1. The passage of the incident
and reflected shock waves cause two schlieren spikes inthe signal, due to steering of the beam off the detector.
After the passage of the reflected shock (marked by a
strong schlieren induced signal), there is a very short
risetime in the absorption signal (�1–5 ls depending
on the temperature and pressure) prior to the steady
absorption due to hot CO2. The absorption cross-sec-
tion is determined during this plateau. This short induc-
tion time is caused by the CO2/Ar V-T relaxationprocess [18,19] which brings the vibrationally cold CO2
to an excited vibrational condition where absorption is
possible. The observation of this induction time pro-
vides evidence that the high-temperature UV absorption
of CO2 comes from a vibrationally excited CO2 ground
state ð1Rþg Þ. It has been shown by Spielfiedel et al. [20]
that molecules with excited bending modes could ac-
count for the hot CO2 absorption at wavelengths longerthan 200 nm because of the significantly better Franck–
Condon overlap these states have with the bent
CO2(1B2) electronically excited upper state. The data
shown in Fig. 1 are at temperatures below 3000 K and
do not show significant thermal decomposition in the
first 200 ls.At higher temperatures (>3000 K) the CO2 undergoes
thermal decomposition and the absorption traces showtemporal decay at long times behind the reflected shock;
an example experiment is shown in Fig. 2. This example
(Fig. 2) shows a significant induction prior to constant
absorption following the incident shock due to vibra-
tional relaxation; the vibrational equilibrated conditions
behind the incident shock are 1812 K and 16.3 kPa.
Although, the vibrational relaxation is resolved after
Fig. 1. Example CO2 absorbance (ln(I0/I)) for experiments at 216.5,
244, 266, and 306 nm.
the incident shock, it is not resolved after the reflected
shock because the higher temperature and pressure
behind the reflected wave have a much faster vibrational
energy transfer rate. Following the passage of thereflected shock wave an incubation period is observed
in the absorption prior to the thermal decomposition
(Fig. 2). The reflected shock conditions are 3838 K
and 83.1 kPa for the experiment in Fig. 2. This incuba-
tion period is somewhat longer than the vibrational
relaxation time and is similar to behavior commonly
seen in shock tube laser-schlieren experiments [21].
CO2 has a large bond energy (125.7 kcal/mol, 43970cm�1) and thus only the high-energy tail of the energy
distribution plays a role in the dissociation, even at the
high temperatures of this study. In the experiments re-
ported here where the average thermal energy is signifi-
cantly below the dissociation energy, thermal
dissociation can only proceed when the highest vibra-
tional levels are populated, and a steady dissociation
rate is not established until the shock-heated CO2
reaches a steady-state vibrational distribution in these
high levels. The incubation period is longer than the
vibrational relaxation time because the time required
for vibrational relaxation in the low-lying vibrational
levels is much shorter than the time required to reach
the steady-state vibrational population in the highest
levels. The incubation period is dependent on collisional
transfer rates from the low-lying vibrational levels at thebottom of the energy ladder to the levels near the disso-
ciation threshold [22]. It should be noted that the
absorption cross-section was determined for these exper-
iments in the incubation plateau prior to dissociation,
and therefore, the cross-section determination does not
M.A. Oehlschlaeger et al. / Chemical Physics Letters 399 (2004) 490–495 493
depend on knowledge of the subsequent chemistry.
Additionally, the experimental temperature during the
plateau is not in question for these high-temperature
experiments because the CO2 vibrational temperature
equilibrates to the translational temperature within
�1 ls (the vibrational relaxation time in these high-tem-perature decomposition experiments). To our knowl-
edge this is the first observation of incubation prior to
dissociation in CO2. We plan to comprehensively exam-
ine this phenomenon in future work.
3. Results and discussion
The measured absorption cross-sections at the four
wavelengths are plotted versus temperature in Fig. 3
and compared with the data from the earlier study from
our laboratory [6]. We find excellent agreement for tem-
peratures between 1800 and 2800 K; at higher tempera-
tures, the extrapolation from the earlier lower
temperature data would predict too large a cross
section.Due to the spectrally smooth nature of the UV CO2
absorption feature [6] the current absorption cross-sec-
tion results in Fig. 3 can be fit to a semi-empirical form
ln rðk; T Þ ¼ aþ bk;
where a = c1 + c2T + c3/T and b = d1 + d2T + d3/T. The
cross-section, r(k,T), is in units of 10�19 cm2 molecule�1,the wavelength, k, is in units of 100 nm, and the temper-
ature, T, is given in units of 1000 K. The following
parameters provide a best fit to the data: c1 = 0.05449,
c2 = 0.13766, c3 = 23.529, d1 = 1.991, d2 = �0.17125,and d3 = �14.694. As is shown in Fig. 3 the scatter of
Fig. 3. Temperature-dependence of the UV CO2 absorption cross-
section. Solid symbols, experimental results: squares, 216.5 nm; circles,
244 nm; triangles, 266 nm; diamonds, 306 nm. Solid lines, semi-
empirical fit to current data; dashed lines, Schulz et al. [6].
the data about the fit is quite small with a 1 � r stand-
ard deviation of 1.8%. It should be noted that this semi-
empirical expression for the absorption cross-section
should not be extrapolated outside the temperature
and wavelength range of the data, but based on experi-
ence from Schulz et al. [6] study interpolation is justified.The absorption cross-section is plotted versus wave-
length in Fig. 4 for selected temperatures below 3050
K to compare with Schulz et al. [6]; this comparison
shows good agreement for temperatures below about
2800 K. The difference between the current measure-
ment and the previous data at 3050 K can be attributed
to over-estimation of the thermal decomposition in the
earlier paper as discussed below.The uncertainty in the CO2 cross-sections obtained in
these experiments is the result of uncertainties in the
experimental temperature and pressure, uncertainties
in the initial CO2 concentration, and uncertainties asso-
ciated with the signal-to-noise of a given experiment.
The experiments in which the smallest absorption was
monitored (low T and long k) provide cross-sections
with the highest uncertainty. The majority of thecross-section results have uncertainties less than 5%;
only a few of the smallest absorption cross-sections
measured at 244 and 266 nm have larger uncertainties,
with the largest uncertainty of 14%.
A comparison of our new cross-section determina-
tions with those of Schulz et al. [6] shows that the agree-
ment is quite good in the middle of the temperature
range, with deviations of 30% and 40%, respectively, atthe highest and lowest temperatures of the Schulz et al.
study which is within the reported uncertainty for those
data at the low and high temperature extremes. At low
temperatures the difference in our findings and Schulz
Fig. 4. Wavelength-dependence of the UV CO2 absorption cross-
section. Solid symbols, semi-empirical fit to current data: squares, 3050
K; circles, 2610 K, diamonds, 2010 K; stars, 1630 K. Solid lines, Schulz
et al. [6].
Fig. 5. Product of cross-section and vibrational partition function
(rQvib) versus inverse temperature (a) and ground state energy versus
wavelength (b). (a) Solid squares, 216.5 nm data; solid circles, 244 nm
data; solid triangles, 266 nm data; solid diamonds, 306 nm data; solid
lines, least-squares fit of form r(k,T)Qvib = reff(k)exp(�e1/T). (b)
Squares, current results; line, linear fit to current results; open circles,
Eremin et al. [24]; ·, Zabelinskii et al. [25].
494 M.A. Oehlschlaeger et al. / Chemical Physics Letters 399 (2004) 490–495
et al. is likely due to limited signal-to-noise; as the noise
in the spectra (Fig. 4) shows, the Schulz et al. results at
low temperatures and long wavelengths are significantly
more uncertain than the current results due to poorer sig-
nal-to-noise. At high temperatures our absorption cross-
sections are lower than those of Schulz et al. due to theiruse of data impacted by CO2 thermal decomposition.
The current experiments indicate that the mechanism
used in the Schulz et al. study (GRIMech 3.0) to account
for decomposition by CO2 has a rate coefficient for
CO2! CO + O that is too fast. Schulz et al. calculated
the average mole fraction of CO2 during their averaging
time period using the GRI mechanism, thus giving a low
CO2 mole fraction. Thus they inferred an absorptioncross-section that is too large. We plan to re-evaluate
the thermal decomposition of CO2 in the future.
The hot CO2 pre-dissociative absorption is a contin-
uum spectral feature caused by the intersection of many
electronic-vibrational-rotational lines, with a dense en-
ergy spectrum typical of a triatomic. If one uses a
quasi-diatomic approximation to describe these optical
transitions [23], it can be shown that the absorbance ata given wavelength, k, is described by
rðk; T Þn ¼ reffðkÞn1 ¼ ½reffðkÞn=Qvib� expð�e1ðkÞ=kT Þ;where the average absorption takes place from a popu-
lation with density n1 at a lower state vibrational energy
e1(k) above the zero-energy ground state and where
reff(k) is an effective cross-section for absorption. Thevibrational partition function, Qvib, is given by
Qvib ¼ ½1� expð�h1=T Þ��1½1� expð�h2=T Þ��2
� ½1� expð�h3=T Þ��1;
where h1 = 1999 K, h2 = 960 K, and h3 = 3383 K. This
description of the temperature-dependent absorption
cross-section allows estimation of the effective vibra-
tional energy of the absorbing lower state, e1(k). Fig.5(a) shows a plot of the product of the measured cross-
section, r(k,T), with the vibrational partition function,
Qvib versus inverse temperature. This plot shows theproduct, r(k,T)Qvib, is linear with a slope that provides
an estimate of the absorbing lower state energy, e1. Thecross-section measurements provide lower state energies
of 1.09 eV (8790 cm�1), 1.48 eV (11940 cm�1), 1.79 eV
(14440 cm�1), and 2.34 eV (18870 cm�1) for absorption
of photons at 216.5, 244, 266, and 306 nm, respectively;
these results can be expressed with
e1ðkÞ ¼ �1:94þ 0:014½1=nm�k eV:
The lower state energy results are compared to previ-
ous determinations made by Eremin et al. [24] (193 nm)
and Zabelinskii et al. [25] (238 and 300 nm) in Fig. 5(b)
with good agreement. Unfortunately, this simplified
photophysical model does not allow for an assignment
of the absorption at a given wavelength to a specificvibrational level(s) because of the complex nature of
the triatomic energy spectrum. Also, note that the line-
arity of r(k,T)Qvib deviates slightly at the highest tem-
peratures (Fig. 5(a)). This deviation might be caused
by enhanced absorption due to transitions to electronic
states other than 1B2 such as 1A2,3A2, and 3B2 that
occurs at the highest temperatures where there is morehighly excited population.
The improved determinations of the CO2 absorption
cross-section presented here should enable absorption-
based thermometry in the high-temperature post-com-
bustion gases in a variety of practical systems. For
example, a two-wavelength temperature sensor using
244 and 266 nm laser radiation in a combustion system
with 10% CO2 at 1 atm, a 10 cm path length, and anuncertainty in absorbance of 0.1% (ln(I0/I) = 0.001),
could be used to measure temperature over the range
of 1900–4500 K with an uncertainty of 6% at 1900 K
M.A. Oehlschlaeger et al. / Chemical Physics Letters 399 (2004) 490–495 495
and 2% at 4500 K. Note, these two wavelengths are
longer than the primary absorption due to aliphatic
hydrocarbons, NO, and O2, and shorter than the OH
absorption feature at 306 nm, and thus avoid interfer-
ence from the major components of combustion effluent.
Such a diagnostic scheme has excellent potential forhigh-temperature gases, especially at high-pressures
such as those found behind detonation waves; this is a
regime where many other schemes fail.
4. Conclusions
The absorption cross-section of shock-heated carbondioxide has been measured at four different laser wave-
lengths (216.5, 244, 266, and 306 nm) in the temperature
range of 1500–4500 K. The high spectral brightness
provided high signal-to-noise ratios and microsecond
time resolution, which allowed accurate measurement
of the absorption cross-section without the use of a chem-
ical model to describe the decomposition of CO2. These
experiments show an incubation period prior to dissocia-tion, observed for the first time in CO2. The temperature
dependence of the CO2 cross-section was used to estimate
the energy levels of the absorbing ground state. It was
found that those molecules absorbing at 216.5, 244,
266, and 306 nm have characteristic vibrational energies
of 1.09, 1.48, 1.79, and 2.34 eV, respectively.
Acknowledgements
This work has been supported by the US Air Force
Office of Scientific Research, Aerospace Sciences Direc-
torate, with Julian Tishkoff as the technical monitor.
The authors thank David Rothamer (Stanford) for dis-
cussions regarding vibrational energy transfer in CO2.
References
[1] H. Okabe, The Photochemistry of Small Molecules, Wiley, New
York, 1978.
[2] M. Koshi, M. Yoshimura, H. Matsui, Chem. Phys. Lett. 176
(1991) 519.
[3] R.J. Jensen, R.D. Guettler, J.L. Lyman, Chem. Phys. Lett. 277
(1997) 356.
[4] M. Ogawa, J. Chem. Phys. 54 (1971) 2550.
[5] B.R. Lewis, J.H. Carver, J. Quant. Spectrosc. Radiat. Transfer 30
(1983) 297.
[6] C. Schulz, J.D. Koch, D.F. Davidson, J.B. Jeffries, R.K. Hanson,
Chem. Phys. Lett. 355 (2002) 82.
[7] J.B. Jeffries, C. Schulz, D.W. Mattison, M.A. Oehlschlaeger,
W.G. Bessler, T. Lee, D.F. Davidson, R.K. Hanson, Proc.
Combust. Inst. 30 (2004) 1545.
[8] N.A. Generalov, S.A. Losev, V.A. Maksimenko, Dokl. Akad.
Nauk. SSSR 150 (1963) 839.
[9] M. Ibreighith, P. Roth, Forsch. Ing.-Wes. 46 (1980)
173.
[10] K.T. Hartinger, S. Nord, P.B. Monkhouse, Appl. Phys. B 70
(2000) 133.
[11] T. Joutsenoja, A. D�Anna, A. D�Alessio, M.I. Nazzaro, Appl.
Spectrosc. 55 (2001) 130.
[12] T. Lee, W.G. Bessler, C. Schulz, M. Patel, J.B. Jeffries, R.K.
Hanson, Appl. Phys. B 79 (2004) 427.
[13] W.G. Bessler, C. Schulz, T. Lee, J.B. Jeffries, R.K. Hanson,
Chem. Phys. Lett. 375 (2003) 344.
[14] D.W. Mattison, M.A. Oehlschlaeger, C.I. Morris, Z.C. Owens,
E.A. Barbour, J.B. Jeffries, R.K. Hanson, Proc. Combust. Inst. 30
(2004) 2578.
[15] M.A. Oehlschlaeger, D.F. Davidson, R.K. Hanson, J. Phys.
Chem. A 108 (2004) 4247.
[16] J.T Herbon, Ph.D. Thesis, Stanford University, Stanford, CA
(2004).
[17] J.T. Herbon, R.K. Hanson, D.M. Golden, C.T. Bowman, Proc.
Combust. Inst. 29 (2002) 1201.
[18] C.J.S.M. Simpson, T.R.D. Chandler, A.C. Strawson, J. Chem.
Phys. 51 (1969) 2214.
[19] G. Kamimoto, H. Matsui, J. Chem. Phys. 53 (1970)
3990.
[20] A. Spielfiedel, N. Feautrier, C. Cossart-Magos, G. Chambaud, P.
Rosmus, H.J. Werner, P. Botschwina, J. Chem. Phys. 97 (1992)
8382.
[21] J.H. Kiefer, S.S. Kumaran, S. Sundaram, J. Chem. Phys. 99
(1993) 3531.
[22] J.E. Dove, J. Troe, Chem. Phys. 35 (1978) 1.
[23] A.V. Eremin, V.S. Ziborov, V.V. Shumova, P. Roth, Chem.
Phys. Rep. (Transl. Khim. Fiz.) 17 (1988) 1275.
[24] A.V. Eremin, V.S. Ziborov, I.M. Naboko, Sov. Opt. Spektorsk.
67 (1989) 562.
[25] I.E. Zabelinskii, I.S. Zaslonko, L.B. Ibragimova, Y.K. Muko-
seev, S.V. Slinkin, O.P. Shatoalov, Zh. Prik. Specktrosk. 4
(1985) 1466.